Equally, verification happens not so much by observing random sense-data to see whether they fit with the hypothesis, but by devising means, precisely on the basis of the larger stories (including the hypothesis itself), to ask specific questions about specific aspects of the hypothesis.
